I'm just gonna post the image of the reconstructions bc I can't in message 💀
btw I got a 16.70 average (idk why I'm better under pressure than normally)
#First official 3x3 average (normal average: 18-19s)
austere magnet
also i cant post youtube vid
💀
lofty sparrow
That’s because you need to be level 20 to send links
You can DM it to me and I’ll post it here
austere magnet
ok the
lofty sparrow
It seems like you don't always plan your whole cross, which leads to terrible midtakes and inefficient solutions. For example in solve 2 you clearly did not plan the orange cross piece for which you then did a y2 rotation to insert it, while you should (almost) never rotate during cross (which is also why I'm not a fan of your xcross in solve 1, as while it was very creative it wasn't that good with 13 moves, including a d and F2, and a rotation). In solve 3 you completely neglected the red piece which was ready to be inserted in one move and broke it while inserting the blue one, which made your solution much longer.
In F2L, I think your main problem is recognition and doing too many rotations (these two are actually connected, because you tend to rotate when you can't find any pieces, which is a bad habit). For example there was a case that you got twice in these solves, which was that there was an edge misoriented in the back slot and a corner with the white sticker facing top. Both times you did a rotation and then paired the pieces up and had to rotate again to insert them, which if you only took a moment to try to recognize the case without rotating, you could have saved a rotation (which maybe doesn't seem too much compared to the time you'd spend on recognizing the case, but if you practice this you'll get better at it, and rotating to find pieces is a bad habit)
Sometimes you also rotate for pair which are easy to solve rotationless and you end up doing two rotations while you could have done none. Like for the 2nd pair in solve 1 for example. Here EO recognition is crucial, which is based on one simple rule: if the top color of the edge matches the color of the front or back face, you can solve the pair rotationless (and moreover with RU or LU moves only)